« Reframing leadership – Seven would-be shifts in how we think and talk about leadership practice | Main | Fundamental flaw in performance-related pay »


Feed You can follow this conversation by subscribing to the comment feed for this post.

John Tropea

Great stuff Chris...this post echoes a lot of your past thinking...

What's your thoughts on this part (just ignore the systems part if thinking in the context of organisations):

"...there may be self-organizing systems without emergent properties..."



Hi Chris,

I like your post and the emphasis on emergence from interactions/conversations.

You might be interested in my InfoQ article - Social Collaboration + Lean Integration = Agile http://www.infoq.com/articles/social-lean-agile

In short, I think the challenge of our time is real-time interaction processing to support feedback, decision support and adaptation.


Chris Rodgers

Thanks for your comment, John.

The person to whom Appelo refers, in making the comment that there may be self-organizing systems without emergent properties, is Peter Corning. His work has been primarily in the areas of the natural sciences (as a biologist) and evolutionary studies, rather than in organizational dynamics.

As I have argued elsewhere in this blog (here for example: http://informalcoalitions.typepad.com/informal_coalitions/2009/08/on-selforganization-and-emergence-1.html) organizations are not living systems. As a result, drawing examples from physics, chemistry and biology to discuss self-organization and emergence doesn't work for me in relation to the dynamics of organizations.

So I'm afraid that, in an organizational context at least, I can see no situation in which a self-organizing 'system' has no emergent properties. As you're aware, I talk of organizations not as systems but (interchangeably) as either "dynamic networks of self-organizing conversations" or as "complex social processes" (of people interacting together). As such, self-organization is the fundamental dynamic of organization, out of which outcomes perpetually emerge. Some of these outcomes (eventually) enter the formal arenas of the organization in the form of structures, systems and procedures, etc. Others remain 'in the shadows', powerfully influencing ongoing interactions but without being formally acknowledged or, in some cases, without conscious awareness.

I hope this helps,


Chris Rodgers

Hi Dave,

Thank you for your comment and reference to your own article. I'm glad that my remarks on the centrality of "2-way conversation" to organizational dynamics and the outcomes that emerge resonate with your own thinking, from a 'Lean' and 'Agile' perspective. It's good to recognize that it's not all about applying the 'right' tools and techniques!

Cheers, Chris

Paul Z Jackson

Thanks Chris, a very clear contribution that may help to keep discussions about complex systems (and emergence and self-organisation) as simple as possible.

John Tropea

Perfect Chris...I wonder then in terms of "systems" whether "...there may be self-organizing systems without emergent properties..."

...just out of curiosity

Maybe Twitter would be a good place to crowdsource such a question.

Chris Rodgers

Hi John,

I struggle to see how there could be a so-called "self-organizing system" which did not have emergent ('system-wide') properties. I guess it might - again - be a question of definition.

As I've suggested earlier, I would have to exclude organizational dynamics from any consideration of this because (a) these are unavoidably self-organizing and generate emergent outcomes; and (b) organizations are not systems.

In one of his papers, Corning talks about a pile of automobile parts which have been thrown together in a heap, without this resulting in any emergent outcome (a car). At a stretch, I guess, these might be thought of as having 'self-organized' into an amorphous pile (although someone did the throwing!). Only when these components are deliberately assembled into a car "in a precise (purposeful) way" does something useful 'emerge', which is more than the sum of the individual parts. I don't know whether this is what Corning intended to convey or if it's what Appelo means in relation to your question.

In any event, I would strongly take issue with Corning's statement in the same paper that "Living systems and human organizations are largely shaped by “instructions” (functional information) and by cybernetic control processes. They are not, for the most part, self-ordered; they are predominately organized by processes that are “purposeful” (teleonomic) in nature and that rely on “control information.” As regards organizations at least, I would suggest that this reflects a very limited understanding of what's actually going on.

As you suggest, Twitter might provide some examples of self-organizing systems without emergent properties. But I'm not sure what that knowledge might satisy - other than curiosity.


John Tropea

@chris yep, I thought as much, it's hardly worth mentioning...I was prepared to be surprised, but thought it was a weird statement.

I applaud snowden for repeatedly making the difference between systems thinking and complexity

“In the idealistic approach, the leaders of an organization set out an ideal future state that they wish to achieve, identify the gap between the ideal and their perception of the present, and seek to close it...Naturalistic approaches, by contrast, seek to understand a sufficiency of the present in order to act to stimulate evolution of the system. Once such stimulation is made, monitoring of emergent patterns becomes a critical activity so that desired patterns can be supported and undesired patterns disrupted. The organization thus evolves to a future that was unknowable in advance, but is more contextually appropriate when discovered."
- Dave Snoweden

hey @dave great article, I really liked it...I have not read much on agile and lean so thanks for this

Dave Snowden has got together with some lean agile guys

Chris Rodgers

Thanks, Paul.

I'm glad that the post makes sense to you.

I think that there is often a tendency to seek too complicated an explanation of the complex dynamics of organization. It's the dynamics that are complex, not the explanation.

Cheers, Chris

Chris Rodgers


Thanks for the Dave Snowden link. Interesting and informative as always.

I agree with his view that there is a mismatch between the 'gap-closing' approach to organizational interventions adopted by mainstream 'systems thinkers' and the complex dynamics of organizations. It's interesting, though, that he describes organizations as complex adaptive systems - which are the externally programmed products of laboratory experiments - rather than using these to draw insights into the unprogrammed dynamics of the political, ideological, and socially complex world of 'real-life' organizations.

In the extract you quote, I found some of his language interesting, too. For example, he talks of "monitoring" patterns - as opposed, say, to paying attention to the repetitive (and novel) themes that are emerging. The former suggests more of an ability to 'stand outside' the interactions (as the programmer does in relation to the design and operation of complex adaptive systems, for example). And this doesn't fit with the recognition that managers (and consultants) are active participants in the ongoing interactional process; not external, objective observers of other people's actions. There is also, for me, too much of a sense that the desired ("more contextually appropriate") outcomes will be achieved as intended, provided that the prescribed approach is followed. From my perspective, managers and others can act with intent but can have no certainty as to what will emerge - however well informed their actions might be.

In the linked post again, Snowden suggests that Ralph Stacey is wholly dismissive of systems thinking ("positively vitriolic" is the phrase he uses in a comment). In fact, Stacey only takes issue with those who see organizations as systems. He has no problem at all with those (like Seddon) who seek to use 'systems thinking' to improve the formal management, operating and information systems through which work is meant to be accomplished.

It is true that Stacey is particularly dismissive of the use of tools and techniques. He sees these as too often diverting attention away from the richness of what's happening in the moment of people's interactions and replacing this with a generalized, supposedly 'best practice' prescription.

Rightly in my view Stacey identifies prescriptive tools and techniques (of the ‘how to change the world in eight easy steps’ kind) as inadequate and misleading abstractions from the complex social dynamics of everyday interaction. At the same time, I believe that helping people to (re-)frame their everyday experiences, issues and events in particular ways is a central aspect of a leader’s task. And it seems perfectly legitimate to me for managers to develop, adopt or adapt sense-making frameworks to stimulate, provoke and facilitate this process. So this is one of the areas in which I depart somewhat from the stance taken by Stacey and his colleagues at the Complexity and Management Centre.


John Tropea

@dave here's a brilliant presentation in getting some more complexity in the lean perspective


Dave Snowden

John T reminded me that my original response had been lost. As I don't have a copy this is a re-creation! We may need to agree to disagree on Ralph. I don't have the book to hand, but I distinctly remember one three page rant (in an otherwise good book) against systems dynamics in which he included Nonaka (correct), Boisot (incorrect and several others. Most CAS people see organisations as systems, the question is what type. In that respect I'm not sure Seddon really fits within any clear theoretical framework. He knows what to condemn and he appears to have the ability to persuade managers to do things better (all of which is to be commended) but all I have read or seen is theory light and largely dependent on his insight/personality.

I think everyone with any sense is against the prescriptive tools etc. I agree with you that Stacy takes this too far, and the essence of Cynefin is to say that we need ontological awareness to use different tools in different contexts.

On pattern recognition - I see nothing wrong with saying that in a complex system you need to monitor patterns, even if you are a part of that pattern you are managing the evolutionary potential of the present. The more complex the system the more you attempt to stand above and aside - which is where our software work lies. For example enabling whole of workforce engagement in situation assessment in real time. Objectivising adductive leads, including the density and diversity of human sensor networks, all increase decision making far more than developing individual capabilities as a direct (note that word) intervention.

Chris Rodgers


Many thanks for your comments.

As you say, we may need to agree to disagree on Ralph Stacey. To begin with, I don’t think that I’ve ever heard or read Stacey rant about anything. It’s true that he argues strongly against many of the established tenets of management orthodoxy and that he is passionate in his belief in what he calls the “complex responsive process” nature of organization. But I think it is stretching it somewhat to call his well-argued challenges a “rant”. I’m not aware of the specific passage to which you refer. However, judging by the names you cite, it’s likely that he was discussing ‘knowledge management’. If so, my guess would be that he would have argued against the ideas that (a) knowledge can be ‘captured’ or managed at all in the ways that this term is usually understood; and (b) that it can be transferred from one person to another in a sender-receiver sort of way. Instead, he would see knowledge emerging in the moment of people’s interactions; and the interactional process itself as a complex responsive, rather than linear transmission, process. But that’s only my speculation.

I mentioned Stacey’s ‘support’ of John Seddon to illustrate the point that it is not the word “system” itself with which he takes issue. His challenge is to those who describe organizations as systems (whether “Complex Adaptive” or any other form), because this runs counter to his complex responsive process view of organizational dynamics (as it does with my informal coalitions/complex social process view). The reality of Seddon’s practice - however he might describe it and whatever the strength or otherwise of its theoretical basis - focuses on improving specific management and information systems. He does this with some success – fuelled, not unreasonably, by his disdain for the use of targets as a performance improvement tool. At the same time, he does not concern himself with the notion of organizations as systems or with the dynamics that flow from that perspective.

Your comment that “the essence of Cynfin is to say that we need ontological awareness to use different tools in different contexts” is, I’m sure, one of the reasons why Stacey would have some difficulty with your position. He would, of course, be able to speak much more articulately on this than I can. The inference I take from your statement is the presumption that managers and others can decide rationally which of the four domains of the framework they happen to be operating in at a particular time and can then use the prescribed approach for that domain to achieve the ends that they desire. Like Stacey, I can’t see this being the case, in the midst of the complex reality of everyday organizational life.

Having just written that statement, it also occurs to me that, paradoxically, the very process of sensing the context, deciding into which category it fits, and then responding accordingly aligns with the suggested response to the “simple” domain in your framework. That is, regardless of the nature of the issue being addressed, it seems that the act of intervention in using the framework itself is presumed to be simple and unproblematic. No doubt you’ll put me right on this.

The notion of managers “attempting to stand above and aside” what’s going on in order to “manage the evolutionary potential of the present” is similarly difficult from my perspective. This is why I raised the original query about your use of the word “monitor”. For me, managers are unavoidably participants in what’s going on – even when they are not present. So participation is their primary ‘state of being’ – whether active or otherwise. They are, as Stacey (after Bourdieu) would say, “immersed in the game”.
They might well strive to adopt a more reflective and reflexive approach to their own and others’ participation in the moment of their interaction – paying attention to the movement of their joint sense-making with others; identifying and exploring dominant themes; challenging clichéd responses (in the sense of actions as well as words); and so on. This is what I would see as the essence of participation from a complex social process perspective. But none of this takes place in any way ‘outside’ those interactions.

And finally, as regards “managing the evolutionary potential of the present”, I would see everyone in the organization as seeking to do this, not just managers. By this I mean that everyone seeks to bring about outcomes that reflect, in whole or in part, their own interests, ideologies and identity. Power relations might ordinarily privilege managers’ participation in this process. But outcomes nevertheless emerge from the widespread interplay of everyone’s interactions, not just those (indeed not primarily those) in which managers are directly involved.

Cheers, Chris

Dave Snowden

Ralph is normally charm personified, but his writing is at times polemical (mine is too by the way) and the passage I mentioned is poor (hence my use of rant). As it happens your description of KM is one that would have been supported by Boisot who Stacy lumps in with Nonaka. My comment on that one therefore stands.

On Cynefin, if you read some of my articles you would find that properly used it is an emergent property of conversations not the result of a "rational" process in the sense you use the words. The fifth domain (you make a common mistake by saying it has four) is key to this and covers ambiguity and inauthenticity. However it critically argues that humans can create order while ants/birds cannot. The framework is defined by contextual narrative that allows matching responses to context. I think you repeat the categorisation error here, but as I say is common when people just look at the model with reading the supporting material and methods so no problem there. I keep trying to think of ways to improve the visualisation to avoid that.

Of course managers are immersed in the game (and I drew heavily on Bordieu by the way), but that does not mean that they cannot increase the level of detachment. You (rather like Stacy) seem to posit an either/or here which I think is a mistake. The fact that you cannot fully detach yourself, does not mean that you cannot observe parallel safe-to-fail experiments. Our approach there is to allow investment in any coherent experiment then see what results, so the manager does not make a choice for an approach, but decides to allow coherent but contradictory approaches to improve understanding through interaction with the system. I think this is about process rather than trying to change the individual qualities of people by the way but that is another subject.

Fully agree that all parts of the organisation should be involved in managing the evolutionary potential of the present. The fact that i say managers should do it does not imply that they are the only ones. I think I used that phrase in my original comment which was lost - do you still have that?

Chris Rodgers

Thanks again for your comment, Dave.

Given the interesting subject matter, the following response is more post-length than comment-length. But, in the words of Mark Twain, I didn’t have time to write a short one!

To begin with, I empathize with your attempts to ‘get the drawing right’. As regards the Cynefin framework, the picture (when drawn with five domains, ‘roughly sketched’ boundaries, and the ‘fold’ between “Simple” and “Chaotic”) seems to me to be entirely consistent with your verbal explanation. But, beyond the drawing, I do still have some queries/issues re its application. By way of reference, I’ve used your co-authored HBR paper of 2007 as an authoritative summary of the framework and its intended use.

I am aware that you advocate a conversational approach to construct the framework in a particular setting. As I understand it, this allows the boundaries between the five domains to emerge through dialogue, based on people’s perceptions of their own circumstances and history, etc. However, in using the resulting framework to guide their approach to their current situation, people still have to decide in which of the domains they consider themselves to be operating at a particular time. From my perspective, I would question the presumed ability of people to do this in the complex social process that is everyday organization. It was this aspect of the Cynefin process in particular that I felt reflected rational assumptions about managers’ capacity to act.

And that was also why I pointed out what I saw as a disconnect between the framework, which acknowledges the complex reality of organizational life, and a process for using it (once constructed) which is more aligned to the guidance you provide for the “Simple” domain – i.e. sense the nature of the context, decide which category this fits into, and respond accordingly. I know that you challenge the use of the term category here. But, despite the socially constructed nature of the framework (when used as you intend it to be used), such construction only seems to relate to the relative positioning of the boundaries. The overall structure of the framework is not up for debate. That is to say, the five domains/categories and their relative orientation to each other are a given.

By the way, I see nothing wrong with this in relation to the potential utility of the framework. My issue is with the implied ability of people to use this to ‘navigate’ their way through the everyday complexities of organizational life. If, on the other hand, Cynefin is used as a way of stimulating and facilitating meaningful, sense-making-cum-action-taking conversations around specific events, issues or challenges (as in the examples in your HBR paper), that is a different matter. But these are the ‘set piece’ events of organization; analogous to the corners, free-kicks and penalties in football (or should I say line-outs, scrums and penalty kicks?). I’m concerned as much with the dynamics of ‘open play’, which accounts for most of the activity. After all, it’s out of the open-play interactions that the above set-piece events - as well as point-scoring, victories and defeats - emerge.

I am aware that there are five domains, not four. However, you say in your HBR paper, “The very nature of the fifth context—disorder—makes it particularly difficult to recognize when one is in it. Here, multiple perspectives jostle for prominence, factional leaders argue with one another, and cacophony rules. The way out of this realm is to break down the situation into constituent parts and assign each to one of the other four realms. Leaders can then make decisions and intervene in contextually appropriate ways.” So this fifth domain appears merely to be a ‘holding camp’ between one or more of the other four. The last two sentences in this extract also add to my unease about the presumed ability of people to deal with this “cacophony” in the analytical way prescribed. In the paper, you also seem to equate political dynamics with this category of “disorder”, seeing it (a) as wholly negative, and (b) as resolvable in the relatively straightforward way described. But I see organizations (that is, people in interaction) as unavoidably political, both in the ‘macro’ sense and also in terms of the politics inherent in everyday organizational interaction.

This points to a more fundamental concern that I would have with the framework, if it were to be positioned as a description of organizational dynamics ‘in the round’, rather than as a tool to facilitate joint problem solving around specific issues. This is because it explicitly states that there are organizational situations which are not complex (that is, either “simple” or “complicated”). I would argue that all situations that involve people are complex, regardless of how ‘simple’ the context might appear to be. The basic unit of organization is the conversation, and this has all of the attributes of complexity that you mention in the article and others that you don’t. For example, conversations exhibit the dynamics of self-organization, co-creation and emergence. Interestingly, column 3 in the table on page 7 of your HBR paper lays bare this inherent complexity, by identifying a number of distinctly human characteristics (i.e. complexities) that can creep into what appear to be even the simplest of situations – and these only relate to things that leaders might do wrong. Those involved in the multitude of conversations and interactions that make up everyday organizational life similarly display all of the frailties, idiosyncrasies and diverse motivations that are the ‘stuff’ of being human. And this, too, means that nothing is simple – or even merely complicated – when it comes to human interaction. It is always complex. People act in a multitude of different ways, at different times and in different circumstances. They behave thoughtlessly, wisely, irrationally, illogically, emotionally, creatively, politically, carelessly, over zealously, caringly, mistakenly, self-centredly, self-sacrificially, etc. So, as soon as you add people to the mix, simple and complicated processes, systems and procedures become complex. And this means that outcomes cannot be predicted or controlled with certainty in any circumstances. Whilst familiar patterns of response might well be the norm, the potential for novelty to emerge and unexpected outcomes to arise always exists.

Finally, I do agree with you that managers (and others for that matter), whilst unavoidably ‘on the pitch, playing’ can, at the same time, seek to take a detached view of what’s going on within and around them. This is what I meant by “…adopt[ing] a more reflective and reflexive approach to their own and others’ participation in the moment of their interaction – paying attention to the movement of their joint sense-making with others; identifying and exploring dominant themes; challenging clichéd responses (in the sense of actions as well as words); and so on.” I see this as a paradoxical relationship between participation and detachment, not an either-or choice. My challenge related to the language that you used, of managers “attempting to stand above and aside” what’s going on. This sounds more like them ‘sitting in the stands’, so to speak, watching other people’s actions rather than actively participating with them in the messy reality of day-to-day life. Here again, this is less problematic if you are thinking about it in relation to a specific issue, rather than as a description of what’s going on in the day-to-day reality of organizational interaction. But, even then, the manager’s actions (including their silence and inaction) will affect they ways in which others’ perceive, interpret and evaluate what’s going on and decide how they will act.

Dave Snowden

Thanks for putting the effort into a reply and I know others are watching this and appreciating the exchange.

Each to their own, but I would be more cautionary than you in taking one article as an authoritative summary of a framework. The HBR article is about the application to Leadership and not about the framework itself or its wider application. I always find it more useful to judge a wider body of work, so just as I would use one article by Stacey to define this work I don't think you should use one article to encompass Cynefin - not that anything in that article is wrong, but its not a complete statement.
I think that people have to define boundaries between different types of things. We do this all the time and it allows us to behave divergently. To say that we make such selections does not mean we are in the Simple domain, that would only apply if the selection method assumed a linear relationship between cause and effect and assumed there was a knowable correct answer. Cynefin domains are defined by multiple mediated narratives within the organisation. They carry with them increasing ambiguity as they move anti-clockwise from Simple. In day to day discourse people use commonly understood examples to create meaning in complex situations (Stacey says something similar as I remember it); so the Cynefin framework builds on that. Oh and yes the overall structure is common, but it emerges from the data. You might want to reflect on language here, without a common language communication is impossible, the same applies to frameworks.

The fifth domain is not a holding domain, that is made very very clear in several articles. Its the normal domain of understanding before people collectively differentiate context to allow divergent actions. That is why it is inauthentic. I didn't think I explained it that badly in the articles, but I will look again to see if there is any way someone could read it as a holding pen, unless of course they were just working from the HBR article without reading the other material in which case I would understand it.

Then we come to the real difference, and here you share a position with Stacy in arguing that all situations involving humans are complex. That is a bigger subject and its one where I think both you and Stacy fail to appreciate the ability of humans to use constraints to create predictability and order. That by the way is no just consciously but also through myths, taboo etc. I worry from time to time that a lot of people are taking complexity theory, but not informing it from anthropology, philosophy, cognitive science and other disciplines.

Oh, and its often more sensible for managers to sit in the stands and limit their engagement ....

Chris Rodgers


I’m very happy to spend time exploring these issues. Thanks to you, too, for continuing the discussion. The following picks up on a number of the points in your latest comment – in particular what you describe as the “real difference” between us.

I didn’t intend to suggest that your HBR article was a comprehensive treatise on the Cynefin framework. That’s why I described it as a summary. And the fact that it was written by you, as the originator and main proponent of the approach, was sufficient in my mind for it to merit the description “authoritative”. Anyway, I accept that you have written extensively on the origin, development and use of the framework; which is clearly held in high regard by a large number of people. My challenge is not on the utility of the framework per se but on the fundamental nature of organizational dynamics.

I recognize the centrality of language in organizations; not least because, in Informal Coalitions, I describe organizations as dynamic networks of self-organizing conversations. But shared language (including the ‘pseudo-language’ of sense-making frameworks and the like) does not of itself ensure effective communication; otherwise we would all be better at it. I think that there is also a question here about utility v universality. If a particular community develops its own language (e.g. jargon) or constructs a framework which ‘works for them’, it can satisfy all of their ‘internal’ communication needs without it having to make sense beyond that community. That is it can emerge from the data without having to conform to a pre-defined structure. If, on the other hand, the latter structure is intended to convey some universal ‘truths’ which apply more broadly than the particular community, then it might make sense for them to organize the data in ways that reflect this understanding. Given that the descriptors of the Cynefin domains and their relative orientation remain intact, it appears to me that it is the latter approach that better describes your use of the framework. As such, I believe that the framework precedes the data, rather than emerging from it. But that’s a matter of interpretation …

As regards the nature of the fifth domain, I’m happy to accept your correction, given that my comment is based solely on my reading of your HBR article. I don’t, though, understand what you mean when you describe the contents of this domain as “inauthentic”.

My use of the metaphor of managers being ‘on the pitch, playing’ rather than ‘sitting in the stands’, was meant to counter one of the basic assumptions of conventional management. That is, that managers can somehow remain apart from the fray, as objective observers of other people’s actions. I suggested in my previous comment that they are unavoidably participants in the ongoing interactional process – even when they are not physically present. At the same time, I acknowledged that their role requires them to strive to achieve a sort of ‘detached involvement’ – to the extent that they can. But, more than this, it’s important for managers to understand that everything that they say and do – as well as everything they don’t say and don’t do – ‘sends messages’ to people. What those ‘messages’ might mean, though, and how they affect people’s actions, are determined not by the managers themselves but through the local sense-making-cum-action-taking conversations that these evoke. So, even where it might appear sensible for managers to “limit their engagement” in a particular endeavour, it can never be as limited as they might imagine or desire it to be.

We now come - as you say - to the real difference in our two perspectives. And, as you are equally correct in saying, I share a position with Ralph Stacey in maintaining that complexity is inherent in all human interaction. I’m not here to speak for Stacey who, as I’ve said previously, would be much more authoritative than I could be in arguing his particular position. However, your comment that Stacey “fails to appreciate the ability of humans to use constraints to create predictability and order” is a gross misreading of his position – and, by inference, of mine. In my own case (which is based on an eclectic mix of perspectives on organizational dynamics – as well as some that are not ordinarily seen as such) I consistently argue that individuals are both enabled and constrained through their interactions with others. It is through these interactions that the constraints to which you refer actually arise. And it is through the widespread interplay of such interactions (both within and beyond the formal boundaries of the organization) that outcomes actually emerge – including the social construction of conditions that might be described as “order”.

In essence, people get together and ‘make things up’. That is to say, through their everyday conversational interactions (both formal and informal), they perceive, interpret and evaluate their experiences and decide – individually and collectively - how they will act. Some of the outputs from these interactions enter the formal arenas of the organization in the form of structures, systems, processes, procedures, or other ‘trappings’ of organization – perhaps even a Cynefin framework. Others remain ‘in the shadows’. These shadow themes significantly affect ongoing interactions and overall outcomes, but in ways that are either known but not acknowledged in the formal arenas of the organization or else beyond people’s immediate awareness. The latter take the form of generalized tendencies to act in certain ways in certain circumstances. These tendencies both affect ongoing interactions and, at the same time, are affected by them. The asymmetrical nature of self-organizing, patterning processes accounts for the sense of ‘order’ that this creates. That is to say, the capacity always exists for new patterns to emerge through the conversational process, but the likelihood is that existing patterns will persist and be strengthened further. This tendency towards continuity (rather, I would say, than “order”) enables people to go on together, without having to think afresh each day what things mean and what they should be doing. This is the essence, from a complex social process perspective, of organizational culture: The more that people make sense of things in particular ways, the more likely they are to continue to make similar sense in similar circumstances going forward. Amongst these ‘shadow themes’ would be the myths and taboos to which you refer. But these do not arise by design; they, too, are emergent outcomes of the complex social process of everyday interaction.

So, yes, despite the intention that people might have to create order by design, and the self-organizing, patterning process inherent in organizational interactions that create expectancy, I would argue that nothing is merely simple or even complicated in relation to the everyday social dynamics of organization. Some of the formally designed structures and systems etc might well be simple in concept or (un)necessarily complicated. But their use by people makes the resulting process unavoidably complex - as are the dynamics through which these designs emerged in the first place and through which they are sustained, developed or ‘corrupted’ by practice. The conversational-cum-behavioural patterns that emerge in practice might tend to be repetitive and routine, giving a sense of continuity or order. Or they might be frenetic (and chaotic), lacking real meaning and engagement, with many disparate actions taking place simultaneously, as participants fail to connect - either with each other or with the ostensible purpose of their interactions. But, in either case – as in the sought-after conditions of free-flowing, organizationally enhancing interaction – these are the ‘product’ of the complex social process of everyday organizational interaction.

Using the Cynefin framework heuristically, to help make sense of a seemingly intractable issue or problem by treating aspects of the situation ‘as if’ they can be thought of as Simple or Complicated, is one thing. But to deny the underlying and inherent complexity of organizational dynamics is quite another. As I said previously, you actually identify (in the HBR article, I’m afraid) a number of these complexities that might undermine the prescribed strategies for dealing with each of the domains. But you seem reluctant to acknowledge these as evidence of the socially complex nature of human interaction.

Dave Snowden

Thanks for that reply Chris, I feel we are moving on to what is one of the most important debates in the application of complexity theory to human systems. It may be getting past a written stage and need a conversation as well - if we can get a conference organiser interested that conversation might also be paid for!

So to respond

  • I think we have resolved the issues over the Cynefin framework and you are correct to say that the overall structure reflects a wider truth. The division of order, complex and chaotic is well established in the literature, the division of order into complicated and simple has been developed in parallel/sequence by a range of people. Disorder I think is mine. The social construction (not to be confused with social constructivism) of the framework is contrasted with the model being pre-given and people assigning data to that. Hence my distinction between a categorisation framework, where the model is given and people fill in the boxes, and a sense making framework where people define spaces based on ordinary language so that the boundaries emerge which is the case with Cynefin. I think that is a substantial and meaningful distinction that means Cynefin avoids your original suggestion that it was a creation of the simple domain.
  • The issue of inauthenticity is linked to the above. It's worth remembering that the underpinning philosophy of Cynefin is in the naturalising tradition of both ontology and epistemology. It therefore works with what is known (or at least coherent if not fully known) from the natural sciences. As such not understanding the ontology or ontologies of a situation (order-complex-chaos) is inauthentic to reality. I would argue that to allow a framework to simply emerge from people's then current conversations would in the main place people in this inauthentic position and thus radically restrict their ability to act in an authentic manner. So yes, natural science is a given, social construction is overlaid on that.
  • We are fully agreed that managers need to be aware that everything they say and do is a part of the system and changes it. I also think that attempts to get them to change their behaviour to be more detached is a mistake, it simply will not happen. However I do think it is possible to create process that forces a greater degree of detachment. We are working for example on whole of workforce engagement in situational assessment, where no individual is aware per se of the whole, but makes a rapid mathematical assessment of an aspect of the situation in parallel with others. This creates a new type of evidence based approach that increases detachment to an extent regardless of the quality of the individual. I have a basic belief, drawn from military work, that we need processes that mean (trivialising a bit) that poor quality people can do the right thing in practice due to training, ritual and process from which individual change follows, rather than focusing on individual change and then hoping the system will sort itself out later.
  • I think if you argue that complexity is inherent in all human action, then you are failing to appreciate one of the key aspects of human systems as opposed to those in nature, namely the ability through constraints etc. to create order in the sense of predictable cause and effect relationships. We are agreed that individuals (and groups, too much emphasis is placed on individuals as agents, mostly its groups and myths which have high agency) are enabled and constrained by interaction. The difference is the degree to which those interactions can create order. Some is emergent, some is designed (and we are able to do that). My characterisation of this I am sure can be improved, but I don't think its a "gross misrepresentation" its a key difference.
  • So I can agree with all of your paragraph which begins "In essence, people get together and 'make things up' but I do not think it is the whole picture. All design is enabled by history, but it is design and it does create predictability (and that for me is the heart of this). While a complex system is always dispositional not causal, an ordered one is causal and thus predictable Humans are able to create the latter, sometimes emergently, sometimes by design, but once done it is stable, repeatable etc. etc. and thus, whatever its origins is ordered within boundaries. I suspect there is a difference as well here between a focus on the individual and individual conversations and one on collective identity - between social atomism and constructivism. But that is a suspicion only, however it would explain some of the differences.

Overall people need structures to create differentiated approaches to both situational assessment and action planning. Without boundaries that allow that it is difficult to move forwards. Where the boundaries come from (in my case social construction of Cynefin) is one thing, the basis of those boundaries (in my case in the main natural science) another. Whatever it requires a recognition that humans are unique in their ability to create predictability, and thus have learnt how, in certain cases, to create order and sustain it.

Johnnie Moore

Thanks to everyone for a very interesting discussion. There are so many nuances to this and it would be easy for me to misunderstand what's being said.

On the other hand, I might learn something if I expose my ignorance.

I'm intrigued by Dave's comment, that "without a common language communication is impossible, the same applies to frameworks." I'm reminded about that old line about England and America being two cultures separated by a common tongue. Reading the very careful debate here, it seems that almost every term can mean slightly different things to different people. So I'm not sure there can ever really be a truly common language.

When I observe humans in conflict, a huge amount of the challenge is recognising that what the sender says and the receiver receives often at variance. Actually, I think the amazing thing about human collaboration is not that there is a common language but there is a capacity to go on together with some levels of confusion. I see endless attempts to close the field of complexity by attempting to agree some more precise model - I really like Dave's term "linguistic conformance" to desribe the pitfalls of this.

I'm a bit sceptical about Dave's argument that Cynefin should be seen as a sense-making rather than a categorisation model. Surely as the Cynefin opus gains popularity, more and more people will come to workshops already knowing the model. Workshops containing such people can't be truly said to allow it to emerge from the data; people will be sitting there knowing the domains already. This would become more prevalent over time.

I think this point may relate to Chris' arguments about managers seeing themselves as part of the system, not outside it. If you go into a workshop understanding Cynefin are you really going to claim it truly emerge from the data unconnected with any expectations of your own?

Dave Snowden

I'm not claiming that lack of knowledge of the model is the issue on sense-making v categorisation. My point is that by allowing boundaries between domains to be negotiated post placement of narrative you avoid categorisation. This is especially true as the fifth domain is key in that method of creation, especially for boundary objects.

We can also create the framework through continuous tagging of multiple data sources over time so please don't confine this to workshops.

Overall this is the key point about complexity - you need some constraint to allowing meaning to emerge. If you allow people to create a model afresh each time from data without some structure then you lack coherence over time and loose the capacity to execute. That to my mind is chaos, where any structure would help, but nothing persists.

So you might end up with a spectrum that ranges from (i) a classic two by two categorisation framework, (ii) the Cynefin four points method (with or without prior knowledge of the model) and then (iii) free form anything goes. Personally I think the real choice is between (i) and (ii) which is the distinction I make between categorisation and sense-making. The third falls into the post-modernist heresy ;-)

Interesting on conflict, a large part of that is not about language per se but evidence. Another potential difference between my view and Stacy/Meade. I think that conflict arises in the complex domain not because we don't understand each other, but because the data can support contradictory theories. Conflict resolution then is about parallel safe-to-fail experiments that may contradict each other to allow solutions to emerge through action, not through conversation per se.

Chris Rodgers

Thanks Dave and Johnnie for continuing the exploration.

To begin with your response of yesterday morning, Dave, I don’t agree that we have resolved the issues over the Cynefin framework. I was happy to see it as a tool that people might find useful to help them make sense of their experience and decide how best to respond in a heuristic sort of way. But I didn’t concede that, “the overall structure reflects a wider truth”. After all, my position is that there are no circumstances involving people that can be described as either “Simple” or “Complicated”. I used the phrase in the discussion about the socially constructed nature of the framework, within the confines of a pre-given set of domains. My point is that if you believe that Cynefin reflects a wider ‘truth’, it would make sense to make this explicit beforehand, in the form of the framework.

However, I don’t think that it does reflect such a wider truth. I think that the division of order, complex and chaotic only exists in some people’s conceptions of the social world by inference from the natural world. But there is nothing in the natural world that reflects such shadow-side dynamics as ideological decision-making, identity formation, informal processes, political behaviour, cultural patterning, diverse aspirations and so on. So I don’t agree that the principles of natural science are a given, in relation to the dynamics of organization.

From my perspective, as I have earlier argued, all situations involving people are inherently and unavoidably complex. It follows from this that there are no ontological differences between the socially complex situations that comprise everyday organizational experience. They are all emergent outcomes of the complex social process of people interacting with each other. These interactions occur directly betwen people; via texts of various kinds; and in relationship with what I have called elsewhere the “imprints of past conversations”. The last of these include the designed structures, systems and procedures etc of the formal organization, as well as the tendencies for people to act in certain ways in certain situations that have arisen through past sensemaking. And so, there is nothing inauthentic in people failing to differentiate situations in line with your labels or in constructing from scratch a framework or other ‘sensemaking tool’ of their own. If people author their own collective story and illustrate it with a self-constructed framework that resonates with their own experience, this would be highly authentic to reality as they had constructed it. As such, I think it would be more likely to facilitate rather than inhibit action, in ways that were meaningful to them.

As regards the third point in your morning comment, I don’t think that I explained well enough that “detached involvement” or “detached participation” is the state in which I believe managers are already operating. That is to say, I’m describing what they are already doing – albeit that they often see this as an either-or choice that they need to ‘balance’, rather than, as I would see it, a paradoxical state that they need to embrace all of the time. For a start, they are unavoidably participating in people’s day-to-day sense-making-cum-action-taking, even when they are not there. This is because people - individually and through their interactions with others - perceive, interpret, evaluate and act upon what they see managers saying, doing and not doing. In other words, managers' actions (and inaction) have symbolic value which is taken up in people’s ongoing conversations. At the same time, established practice defines their role primarily as one of standing apart; observing the ‘bigger picture’ so that they can set policies, observe performance and control people’s actions. So detachment and participation are a given. It’s just that managers are often not aware of the impact of their (observed and often informal) words and actions on people's behaviour. Nor, typically, are they actively involved enough in the everyday conversational life of the organization, through which people make sense and take action. This is not least because they are often too detached in their pursuit of the goal-maximizing objectivity sought by conventional management wisdom.

I’ve mentioned above why I think that human processes (I don’t see organizations as systems) are different from those in the natural world. And also why my reasons for saying so differ from yours, Dave. As I have earlier acknowledged, it is true to say that human beings tend to seek order; putting structures, systems and regulations etc in place to try to bring this about. But, people being people, the outcomes that emerge from the complex interplay of all of this activity cannot be controlled – or predicted with any degree of certainty. Of course human beings are capable of designing technologies that can facilitate their own human being and human doing – such as building a rocket and related technologies capable of putting a man on the moon. But organizations are not technologies. And the dynamics of organizations are not rocket science – they’re much more complex than that.

Because I don’t see organizations as systems or adopt a systems perspective in relation to organizational dynamics overall, I also have a problem with your notion, Dave, that groups and myths have agency. You might think that it is splitting hairs to say that it is individuals in interaction with others (rather than groups per se) who have agency. But, from a complex social process perspective, there is no ontological difference between the individual and the organization (or anything ‘in between’). As Taylor and van Every say “What we call organization is generated in the same conversation where individual actors find their identity.” Similarly with myths. These have themselves arisen from the complex interplay, over time, of many local conversations. A myth might well be drawn upon (consciously or unconsciously) in people’s current sense-making interactions, and so affect the response that they make to a situation. But it is people who act. And only to the extent (and in the ways) that such myths are taken up in present-day, local conversations do they have any meaning. The same can be said for the formal structures, systems and procedures through which managers seek to impose order. They only have any meaning to the extent, and in the ways, that these are taken up in ongoing interactions.

From the above, it is perhaps easy to see why you might think that I have placed an emphasis on the individual and individual conversations as opposed to “collective identity”. This is again positing different ontological levels between the individual and the collective, which I referred to above. The dynamic that I want to draw attention to is people’s ongoing, conversational interactions and the widespread (organization-wide, if you like) interplay of these, out of which ‘collective’ outcomes (including any sense of collective identity) emerge. So I’m interested in the dynamic interplay between ‘local’ and ‘global’ at the same time.

Picking up on the final statement in your comment of yesterday morning, I guess I would see humans as unique in their ability to understand that, as far as the social worlds is concerned, nothing is certain. And this is echoed, I think, in your comment Johnnie, “I think the amazing thing about human collaboration is not that there is a common language but there is a capacity to go on together with some levels of confusion.”

In the sensemaking v categorization point that you both comment upon in the later exchange, I think that you, Dave, are unnecessarily polarizing these two acts. As I understand it, you distinguish between sensemaking, in which a framework emerges from the data, and categorization, in which the data is allocated to pre-defined categories. I prefer to distinguish between prescriptive frameworks and sense-making frameworks. The former tend to define a number of possible states (often, as you say, in the form of a 2x2 grid) and define one of these states as the ideal to which the user should aspire. The latter help people to make sense of their situation in some way or other, one of which might be to organize their understanding according to some pre-defined categories. Language is one such sense-making ‘framework’ in which the categories (words and phrases) precede the experience. Some 2x2 grids and other frameworks might similarly facilitate this sensemaking process.

In your response to Johnnie, Dave, you state that the key point about complexity is that “you need some constraint to allowing meaning to emerge”. I see the interactional process as providing the enabling-constraints through which meaning is co-created locally. I don’t agree at all that the absence of externally imposed constraints means that whatever results necessarily lacks coherence over time or that, as a result, people lose the capacity to execute. In some cases it will do, of course. But that is the nature of emergence. And, as I suggested above, any formal constraints that might be imposed will themselves only have meaning (and therefore impact) to the extent that they are taken up by people in their ongoing interactions.

On a final point ("Thank goodness for that," I hear you say), concerning your exchange on conflict, I think that there are myriad reasons why conflict and contention arises in organizations. Some of these might concern lack of agreement on potential ways forward. And, in such cases, your “safe-to-fail” experiments, Dave, might well be one way forward. Deciding at the end of such experiments which merits further action or exploration is, of course, resolvable only through conversation. Other tensions might be the result, as Johnnie says, of misunderstanding, which is an ever-present possibility in human interaction – and often the source of creative and innovative ways forward. Other conflict arises from the clash of interests, ideologies and identities etc, which arise both from the inherent design of organizations and from people's pursuit of their personal and/or shared agendas.

Dave Snowden

My point was that we had resolved the issues arising from your taking Cynefin from the HBR article alone and that i consequence we were not moving on to the real debate - and one which is not really aired enough.

I think (correct me if I am wrong) that you accept order-complex-chaos reflects a wider truth in physics but not in human systems. Cynefin reflects wider truth but splits order into simple and complicated and adds disorder to account for some of the key differences between complex systems in humans and those elsewhere in nature. I won't repeat the point I made in my reply to Johnny, but that deals with the "explicit before hand" point you make.

We are also agreed that there is no equivalent in the natural world to ideological decision making, identity formation etc. (well there is to an extend in primates and possibly more extensively but its not significant). So I think we are agreed that human complex systems represent a different order of complexity from that found in chemistry and biology. That causes me to doubt much of the modelling work that is done in complexity in universities in part because I don't think humans have a single identity that can be called an agent, or that they make decisions based on rules. For that reason I distinguish cognitive complexity from complexity per se. That puts you, Ralph and myself is a similar camp against many others. However, the way we deal with it is different.

Your statement that "all situations involving people are inherently and unavoidably complex" summarises the position. In contrast with this I think that humans, due to their capacity for ontological awareness, are able to create ordered systems within and through constraints. Thius ranges from highly overt and explicit use of process for cheque payment systems, quality control etc. up too in large populations the evolution of ritual, myth etc. These may start as complex but they create coevolutionary structures that in effect create ordered system. That is to say they exhibit causality and/or deep stable dispositionality that allows for repeatability. Cynefin with its simple and complicated domains reflects this.

An emergent outcome of a complex process (to use your language), can in human systems create order. It does not naturally exist, but we have evolved to create it.

I am also happy to agree that managers are always involved and can also to a degree detach. My wider point is that with modern technology we can increase that detachment, objectifying adductive linkages and therefore take a different approach to evidence based policy. It requires process to do it and large numbers but it is one valid approach to managing complexity. Your later comments seem to indicate some agreement with this.

I agree that people act, but to a degree and we get into complex issues around free will here. In effect Myths, rituals etc can direct human behaviour to the point where there is no agency per se in the human actor, but instead it rests with the myth. Individual interacts with collective but in the main individual meaning is derived from interactions, not the other way round. As Alicia Juarrero says, meaning exists in the interactions between things not the things themselves, and there in lie myths and many other aspects of our collective identity.

I'm happy to accept your distinction between prescriptive and sense-making frameworks by the way, its another way of making the same point. Cynefin for example argues that all the domains have equal validity which fits your sense-making definition. However I do think the precedence of data and framework is important and I have nothing more to add on that to what I have said before.

Otherwise conversation is one, and possibly the dominant way of deciding what we should move forward with, but its not the only way, self-evident results of action are another. I think one of the dangers with Stacey is that he privileges conversation too much. I just came back from a project in the US where we moved executives away from conversation to acts (without conversation) that made negative stories difficult to maintain. Conflict creates more scanning (Klein and Snowden but can't remember the exact reference) than consensus in military decision making and so in. I like Meade, but I think its time to move on.

Chris Rodgers

Thanks, again, Dave.

It seems that there are several things on which we largely agree! Not least, we agree that the ways in which complexity theory is too often imported directly from the physical/ natural world into organizational settings is misleading and unhelpful. I similarly agree that human beings do not have a single identity and that they are not rule-following ‘agents’, as implied by unalloyed Complex Adaptive System thinking. As you say, this places us on the opposite side of the fence from most others who write and talk in this area.

Clearly, we each approach the topic from different intellectual and practical standpoints and ‘rub up against each other’ in relation to some of our detailed explanations and the implications of these for everyday practice. For my part, I’m trying to help ‘shift the patterns’ of understanding and behaviour away from those advocated by conventional management theorists – including those who use the language of complexity to repackage old prescriptions. As Keith Grint puts it, “Much of what is taught in management or business schools, or written about in management or business books, is a banal paradox. It is banal in that it appears to regurgitate what everyone already takes for granted and knows to be true. It is a paradox because, despite being full of common sense, it doesn’t seem to work.” Seeking to build informal coalitions of support around ideas and practice that challenge what Ralph Stacey would call “the dominant management discourse” is central to this.

Without replaying our earlier conversation, I think that what you describe as a separately definable, ontological state of “order”, I would see as a desire for and sense of order (or continuity). In my view, this emerges from, and is sustained by, the complex social process of everyday interaction. That is, I don’t see it as a mature expression of these dynamics, which started as complex but evolved to become ordered and predictable. It reflects a paradoxical state of order-disorder, predictability-unpredictability, and so on. Many of the designed processes, systems and procedures ‘do what they say on the tin’ - at least most of the time. Many others, designed with similar intent, fail miserably to achieve the desired ends. So a sense of order, predictability and control co-exists, in a paradoxical way, with messiness, uncertainty and unforeseen outcomes. These, from my perspective, are all characteristics of the socially complex dynamics of organization. This is also how I see myths, folklore and the like working – not by acting on people but by people taking them up – or not – in the particular circumstances of their ongoing interactions. The patterning bias towards continuity suggests that action is likely to continue in line with existing ‘mythology’ and unquestioned assumptions of all kinds. But the capacity always exists for sense-making to ‘switch over’ into a different pattern, and for new responses to emerge.

Like you, I believe that meaning emerges from interaction; that is, that it is socially constructed. So we have a degree of agreement here, too. However, I don’t see meaning as existing or being ‘stored’ in any way ‘outside’ people’s interactions – as you seem to be implying in relation to myths and collective identity. So, if that is what Alicia Juarrero means by meaning not existing in “the things themselves,” I would have to disagree with her on that. I see meaning emerging in the moment of interaction – as established patterns of understanding/response are ‘triggered’ (and potentially shifted). This is also the case even if an individual is thinking alone (i.e. talking to his- or herself). Their internal conversation always draws upon broader narratives and the more generalized patterns of assumptions, beliefs and behaviours that reflect the broader communities to which they see themselves belonging. So, in this sense at least, thinking is always a social activity.

I agree wholeheartedly that precedence of data over framework – and, in particular, the meaning of that data - is important. In relation to one of my own sense-making frameworks, a client put it very nicely when she said that working with that particular framework is “not like playing Sudoku”. That is, it is not about ‘filling in’ the boxes as quickly as possible with the ‘right answer’. The value is, as she says, in the conversations that people have - facilitated and provoked by the framework.

On your final point, about the rightness or otherwise of privileging conversation, I have to disagree strongly with your suggestion that Ralph Stacey’s approach – and, again by inference, my own – places too much emphasis on the role of conversation. I first settled upon the notion of informal coalitions as a key dynamic of organizational change from my observations and reflections as an in-house manager in the mid-90s. I saw language and everyday conversation as central to this. Although I have studied and admired Stacey’s work for around 20 years now, I arrived at my view of organizational dynamics independently – spurred further by my work in the late-90s on a Masters course. So my thinking was not influenced by an understanding of Meade. Or Elias. Or any of the others to whom Stacey refers. I am much more likely to draw inferences from the early writing of de Bono on the ‘mechanism’ of mind than from Damasio. So my perspective was derived much more from my practical experience as a manager than as an academic and radical thinker on complexity. Indeed, one of the things that pleased me most was when someone once said that I manage to talk about complexity without talking about complexity! And that is, no doubt, why my view differs from that of the Stacey school ‘around the edges’. They would undoubtedly highlight differences between their complex responsive process perspective and informal coalitions.

Fundamentally, though, I think that Stacey has got it right. As I said in an earlier comment, a conversation is the fundamental unit of organization. And it follows from this that conversational dynamics are the essence of organizational dynamics. On that point, though, I’m not expecting that we will reach agreement.

Dave Snowden

I think we are now at the heart of the disagreement, and the agreements which is a good place to be! My apologies for assuming that you were (per Stacey) focused on Meade. Its probably time to move the discussion to the physical world if we can.

Some reflections arising from thison my own blog.

Chris Rodgers

No problem with your assumption re Meade. But I'm just a simple soul. Ooops! I mean complex, of course.

Thanks again for the stimulating conversation.

Roy Williams

First, many thanks for a fascinating conversation.

Second, the nub of this, for me, is based on Dave's point that

"•I think if you argue that complexity is inherent in all human action, then you are failing to appreciate one of the key aspects of human systems as opposed to those in nature, namely the ability through constraints etc. to create order in the sense of predictable cause and effect relationships"

I go along, largely, with the Cynefin framework. I make the distinctions in slightly different terms: the key for me is the difference between variables 'with or without attitude', by which I mean, simple, self-organising, reproducing, and multiplying capability.

It might be useful to divide the problem, for starters, into three scenarios: those in which, with a some effort, you can achieve predictability, compliance, and participation.

1. Predictability. This can be achieved if you work exclusively with variables that are not self-organising (to start with, lets limit this to the lack of RNA/DNA), and which can accordingly be treated as part of predictable systems with alacrity. As long as you dont work at too massive or too nano a scale, they behave as predicted (given the benefit of a few hundred years of science). So yes, we (uniquely) can create predictable order.

2. Compliance. This can be achieved in a 'mixed ontology' of variables with and without attitude. You can choose to constrain the variables that you are working with 'as if' were predictable, simply by denying them any self-organisation. Ecology, democracy, global warming, Fascism, Stalinism, Arab Springs, etc, teach us, sadly, that this is unsustainable - eventually.

3. Participation. You can 'govern with consent' as the primers of political science teach us - and it applies to ecologies, Ozone holes and all sorts of other things too.

So yes, most things are complex, some things are (within specifically domains) predictable, and we we are arrogant (and sometimes lucky) enough think we can get away with treating some complex things as if they were dumb and predictable.

We need to know our way around the Cynefin (or a comparable) framework, even if only to know when we are in denial, and when our 'order' is likely to remain predictable, and when it will show its attitude, and is likely to kick back. After that its up to you.

Chris Rodgers

Many thanks for your comment, Roy.

I recognize that you find value in seeking to distinguish between various ‘categories’ of phenomena, in much the same way that Dave Snowden argues via his Cynefin framework. At the same time, I don’t believe that there is anything in your argument that negates my contention that complexity is inherent in all human action and, in particular, inter-action.

I fully understand that, as human beings, we seek to impose order on the complex social dynamics of our everyday lives. The felt need for certainty, predictability and control exerts a strong psychological pull in that direction. And frameworks that offer to resolve the complexities of day-to-day living into clearly defined categories (or domains, as Dave prefers to describe these in the Cynefin framework), each with its own prescribed ways of responding, can be very appealing. These also sit well with the dominant management discourse, which ascribes to managers the ability to assess a situation, choose an appropriate course of action, and deliver the desired outcome. Indeed, if they can’t do so, that same discourse would maintain that they’re clearly not managing! So the contention that “human systems [have] the ability through constraints etc. to create order in the sense of predictable cause and effect relationships” appears commonsensical and is a taken-for-granted assumption of mainstream management practice.

However, to paraphrase Dave’s comment, this perspective fails to appreciate one of the key aspects of human processes (not “human systems”, by the way). That is, that the design, development, acquisition, deployment, management, use, abuse and abandonment etc of the very technologies, systems and procedures that are intended to produce the desired order are themselves emergent outcomes of the complex social process of ongoing human interaction. And this means that, despite the ‘best laid plans’, predictability of outcome cannot be guaranteed, even in those situations that are highly automated or seemingly mundane. Some of the outputs along the way might, for all practical purposes, be largely predictable – “all other things being equal”. But the notion of cause and effect has little meaning when applied to the overall dynamics of organizational performance. There, outcomes are governed as much – if not more – by shadow-side themes arising from informal processes and practices, competing interests and ideologies, issues of identity, shifting power relations and other relationship dynamics, etc as they are by the formally imposed requirements.

Crucially, too, people always act (indeed they can only act) in the ‘here and now’ of the specific situations and relationships that they find themselves in – at that time, in that place and with those people. So grand designs and prescribed actions are always mediated by, or given meaning through, people’s in-the-moment interactions.

You seem to acknowledge this to some extent by reserving what you call “predictability” for those situations involving “exclusively ... those variables which are not self-organising”. In the context of organizations, this is likely to relate to some form of technology in which the pushing of a particular button or the pulling of a particular lever will (all other things being equal, again!) produce a predictable response from some other part of the equipment. Unfortunately (or fortunately, perhaps) when you add people to the mix (i.e. when such technology is made use of by people in an organizational or other social setting), the presumed predictability becomes – well – less predictable!

As regards your “compliance” category, I would ask you to reflect on the complex social process through which the formal constraints on people’s actions arose in the first place. And on those same dynamics through which the outcomes desired are being pursued by those with the relative power to impose their will. Also, even when subjected to such constraints, many of those denied the right to organize formally will nevertheless be seeking to mobilize support around an alternative construction of events. This is evidenced by the examples you cite, in which a sufficiently powerful coalition of support eventually emerges to pursue one or other alternative course.

Then you offer, as a third category, those conditions which you refer to as “participation”. Here again, you seem to be denying the complex social processes inherent in the “govern[ing] with consent” that you see as the touchstone of this approach. Most of the significant aspects of this take place ‘in the shadows’ of the formal structures, systems and processes established to achieve such governance. And, as in the other two categories, the formally established ‘trappings’ of organization are themselves the products of the complex social process of everyday interaction – infused with differing, constantly changing, and frequently competing ideologies, interests and identities.

So, whatever labels might be attached to particular sets of circumstances that emerge, the underlying dynamics are the same in all cases. That is, outcomes emerge from the complex social process of everyday interaction. Or, as I describe it in Informal Coalitions, from the dynamic network of self-organizing conversations that comprise ‘the organization’. As I suggested in an earlier response to Dave Snowden, I see no ontological differences between the supposed ‘levels’ of complexity that are implied by the domains of the Cynefin framework – or in your own classification.

Mireille Jansma

Interesting and enlightening conversation, thanks!

David, I have a question about the Cynefin framework 'emerging' from data. Happen to be rereading David Deutsch' book "The Fabric of Reality", and he makes a very strong case against the idea that we can get to new theories on the basis of observations. But perhaps that's not what you mean?

Thanks again,

Mireille Jansma

Chris Rodgers

Hi Mireille,

Many thanks for your comment. I'm glad that you enjoyed the conversation!

I've drawn Dave's attention to you question about David Deutsch's book.

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Your comment could not be posted. Error type:
Your comment has been saved. Comments are moderated and will not appear until approved by the author. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.


Post a comment

Comments are moderated, and will not appear until the author has approved them.

Your Information

(Name and email address are required. Email address will not be displayed with the comment.)